Fugu AI: what the agent does and how to choose a model

1. What the Fugu agent actually does

Fugu AI is Sakana AI's multi-agent model family, delivered as a single OpenAI-compatible API. Each request is answered by a small team of specialized agents that Fugu coordinates and synthesizes into one response: fugu for speed, fugu-ultra for depth, fugu-max for cost-performance, and fugu-cyber for defensive security.

Fugu is a learned multi-agent system. A request can be answered directly or routed across a small team, then synthesized into one response. The orchestration is learned rather than defined by a workflow you maintain. That is the core trade: less workflow code for you, less control over the path.

Official product copy says Fugu can exclude specific providers or models from its pool, while Ultra and Max use fixed pools. New public frontier models are expected to take roughly two weeks of training and evaluation before a Fugu update. That means the behavior you measured can drift on the vendor's schedule, not yours.

2. The orchestration mechanism, in six constraints

• Fugu coordinates a pool of models through a single OpenAI-compatible (and Anthropic Messages-compatible) interface.
• Fugu allows provider or model opt-outs from its pool; Ultra and Max use fixed pools with no opt-outs.
• Current docs describe Ultra and Max as fixed pools and no longer state an exact agent count.
• Routing details are proprietary and not exposed to the caller.
• Orchestration work produces extra billable tokens on Ultra and Cyber.
• New frontier models take roughly two weeks of training and evaluation before a Fugu update.

If your audit requirement is "name every worker," the answer is a self-managed graph, not a different Fugu temperature. See Fugu versus self-managed orchestration for the decision rules.

3. Orchestration tokens are extra work you pay for

OpenAI-shaped field names make it tempting to treat every token_details entry the same way. The ordinary cached_tokens subset does behave like OpenAI — it is already inside input_tokens — but official Sakana pricing says the orchestration fields are additional real usage billed at the matching input, cached, or output rate.

{
  "input_tokens": 120,
  "output_tokens": 80,
  "total_tokens": 200,
  "input_tokens_details": {
    "cached_tokens": 0,
    "orchestration_input_tokens": 450,
    "orchestration_input_cached_tokens": 100
  },
  "output_tokens_details": {
    "orchestration_output_tokens": 220
  }
}

Each cached field is a subset of its own input total — cached_tokens of input_tokens, orchestration_input_cached_tokens of orchestration_input_tokens — so subtract before applying the full input rate and reject any count that exceeds its input. Only the orchestration input and output fields are extra outside the ordinary request counts.

Ultra's published standard rates on October 14, 2026 are $5 input, $0.50 cached, and $30 output per million tokens, rising above 272K request context to $10 / $1.00 / $45. Fugu Max is flat at $2 input / $0.25 cached / $6 output at any context length, so no threshold adjustment applies to it. The 272K threshold is per request, not a monthly total. Walk the arithmetic in orchestration token accounting.

4. Fugu versus Ultra versus Max versus Cyber: the selection table

Each row is a decision you can actually make before writing a request. Where a row says "not Fugu," the reason is a documented product rule, not a quality ranking.

The model chooser encodes region, routing visibility, and workload into the same rule-based recommendation, including the "do not use Fugu here" outcome.

5. Responses API essentials

Sakana AI recommends the Responses API for new integrations, especially for tools, multimodal input, reasoning controls, and function calls. The request shape is familiar, but several fields do not behave the OpenAI way.

import os
from openai import OpenAI

api_key = os.environ["FUGU_API_KEY"]
base_url = os.environ["FUGU_BASE_URL"].rstrip("/")
if not base_url.endswith("/v1"):
    base_url += "/v1"

client = OpenAI(api_key=api_key, base_url=base_url, timeout=120.0, max_retries=2)

response = client.responses.create(
    model="fugu-ultra",
    instructions="Find correctness risks first. State uncertainty explicitly.",
    input="Review this migration for rollback risks.",
    max_output_tokens=2000,
)
print(response.output_text)
print(response.usage)

Complete client setup and streaming guidance are on the API quickstart.

6. Diagnose failures before you retry

The full failure tree is on the troubleshooting page.

Frequently asked questions

Sources and verification boundary

• Sakana AI Models documentation: IDs, pool behavior, supported fields, and usage shape
• Sakana AI Pricing documentation: Ultra and Max rates and orchestration field billing
• Sakana AI Fugu product page: agent behavior, opt-outs, and update cadence

This page describes documented behavior as of the October 14, 2026 recheck. Recheck the official Models endpoint before a 2026 production release.